EP3411803A1

EP3411803A1 - Device and method for processing a binary-coded structure document

Info

Publication number: EP3411803A1
Application number: EP17715073.7A
Authority: EP
Inventors: Richard Kuntschke; Daniel Peintner; Andreas Scholz
Original assignee: Siemens AG
Current assignee: Siemens AG
Priority date: 2016-04-12
Filing date: 2017-03-28
Publication date: 2018-12-12
Anticipated expiration: 2037-03-28
Also published as: WO2017178222A1; CN108885612B; DE102016206046A1; US11545997B2; EP3411803B1; US20190155875A1; CN108885612A; EP3411803C0

Abstract

In the method according to the invention, a binary-coded structure document, in particular in an EXI format which is based on a structure document designed in a text-based description language, in particular XML, is provided with a binary-coded index document, which is based on a structural part of the structure document. A storage representation of the binary-coded index document is stored in a working storage unit of the device, and the binary-coded structure document is stored in a background storage unit assigned to the device. A processing operation or a request by at least one processing process is carried out by accessing the storage representation of the binary-coded index document via an object interface, and if necessary, fragments of the binary-coded structure document are loaded into the working storage unit by the object interface. By carrying out processing operations on the binary-coded index document, also referred to as XIndex, according to the invention, faster request operations are facilitated in that specific requests can already be processed using the index document BCI without the involvement of the binary-coded structure document.

Description

description

Device and method for processing a binary-coded structure document

The invention relates to a device and a method for processing a binary-coded structure document.

Text-based description languages for the specification of data formats and the methods required for processing the data are known in the prior art. A structural document referred to in this description includes at least in part ^¬ as passages in a description language. A well-known description language is Extensible Markup

Language, "XML abbreviated with which hierarchically struc ^¬ tured data in text form or" be Piain text "described. The description language XML is set once for a plattformun ^¬ dependent exchange of data between computer systems. Due to the textual nature of XML, it is readable ^both by machines and by humans.

Structural documents present in the XML descriptive language are usually used for processing - in particular by searching, filtering, manipulating, transforming, and interrogating data - the data contained therein in a memory representation, which in the professional world is also referred to as an "in-memory" model is known. A known specifics ^¬ tion for accessing a memory representation of XML documents is the "Document Object Model" or DOM, which provides an object interface for data access by one or more machining processes. The object interface, for example, as a programming interface ^¬ or API (Application Programming Interface) executed. However, the memory representation is many times larger than the corresponding structural document. For this reason, it is often not possible or only possible with restrictions to process more extensive structural documents in resource-limited devices-that is, via limited storage space and / or computing power. Such resource-limited devices include, for example, embedded systems or "embedded systems" for monitoring, control or regulating functions, for example in the form of so-called intelligent sensors, in particular in an assembly line or in a vehicle.

Frequently, a fast data exchange between devices is required, which is insufficiently possible with a textual description language such as XML. Therefore binary Re ^¬ presentations of XML have been proposed for efficient data exchange. A as "Efficient XML Interchange" which eclip ^¬ shortens EXI, called binary representation of XML required compared to text-based XML data less transmission bandwidth for data transfer between devices or computer systems of a general nature.

The exchange format EXI does not only develop its advantages in the name-giving data exchange between computer systems, but also in the in-device processing of a binary-coded structure document in resource-limited devices. A device-internal processing of the binary-coded Strukturdo ^¬ document, ie in EXI format, often proves to be more effizi ^¬ enter compared to the use of the XML format, especially since the data between devices are already exchanged according to EXI specifications.

For a device-internal processing, the binary-coded structural document is then - in analogy to the this närkodierten structural document underlying, constructed in a text-based description language Strukturdoku ^¬ ment - transferred to a memory representation to one or more processing processes via a Obj ektschnittstelle access to the data of the binary-coded Strukturdoku ^¬ ment to allow. In-device processing of the binary-coded structure document also encounters limits in resource-limited devices, which require considerations of the size of the memory representation in consideration of a limited storage capacity of a working memory of the device.

Exceeds the memory representation of a document structure ^¬ ments a predetermined size, the corresponding structure has to be cut to an acceptable size document waiving originally scheduled func ^¬ ities. Schei ^¬ det this possibility proprietä ^¬ re solutions for a streamlined memory representation or alternative machining processes need to be taken, inevitably, which can be personalized for a particular application. This inevitably means that an alternative solution has to be developed for new or changed use cases. This is not advantageous for the basic goal of a generic solution.

It has also been shown that read operations in the possible editing operation predominate over write-access operations. In contrast to read operations, in general machining operations, data is not only read out but also changed, written or restructured. In turn, in these read operations, structural read operations on a binary-coded structural document outweigh read operations in which content is to be read. The first mentioned structural In particular, read operations include queries to a hierarchy context, to a number or to specific categories of elements to be determined within the structure of the underlying structure document. These structural read operations usually require a processing of the complete memory representation of a structure document, which is represented in the memory representation eg in a tree structure. In the following, for reasons of generality, all the reading and writing operations mentioned are referred to by the general term machining operations.

The present invention has the object to provide a method in a particular resource-limited device, by which processing operations on a

Memory representation of a binary-coded structure document are made possible without the size of the binary-coded structure document abuts the limits of the resource-limited device.

The object is achieved by a method having the features of patent claim 1.

The invention is borne by the central idea of creating a binary-coded index document, the structure with the structure of binary coded to edit the structure of the document is the same but keeps ent ^¬ in contrast to the full binary coded structure document that not all information, so for example, no attributes, no Wertzuwei- sations of attributes and no content of elements.

Both the document and the binary-coded structure binärko ^¬ all official document index, is based on an identical structure document, which in a text-based description language is constructed, preferably in XML. While the binary coding ^¬ th structure document, the full contents of the document structure is based, is the binary coded index document unmarried ^¬ Lich a structural part of the circumference of the structure document to-reasons. The index document contains a so to speak Binärreprä ^¬ presentation of scarce hierarchical data structure of the elements defined in a tree structure, without the zughörigen content. Because of this merely structural partial extent, the binary-coded index document naturally requires a considerably smaller amount of storage space than the binary-coded structural document.

The process of the invention now provides that a SpeI ^¬ cherrepräsentation of binary coded index document is stored in a memory of the device, while the bi ^¬ närkodierte structure document is stored in a device associated with the background memory.

According to the invention, processing operations are carried out by at least one machining process

Access to the binary - coded index document via an object ^¬ interface, wherein, if necessary, fragments of the binary coded ^¬ th structure document are loaded by the object interface in the main memory.

The object is further achieved by a computer program product for processing the method according to the invention. The computer program is processed in a processor or controller which executes the method with the processing.

The object is further achieved by a device for the particular resource-limited processing of a binary-coded structure document. Further embodiments of the invention are the subject of the dependent claims. In the following, further embodiments and advantages of the invention will be explained in more detail with reference to the drawing. Showing:

1 shows a schematic structural representation of a processing of a binary-coded structure document in a device; and;

2: a schematic representation of a structure of a

Index document in comparison with an underlying structural document constructed in a text-based description language.

FIG. 1 shows, in a non-conclusive manner, a schematic structural representation of a processing of a binary-coded structure document BCS in a device.

In an object interface Ol a Speicherrepräsenta ^¬ tion of a binary coded index BCI document is stored. The object interface Ol allows a respective machining process XP1, XP2, XP3 access to the Speicherrepräsenta ^¬ tion of binary coded index document BCI. The presentation of a ^¬ Speicherre binary coded index BCI document and the object interface Ol is a - assigned to the device's memory - not shown.

In a - not shown - background memory of the device is a binary-coded structure document BCS behind ^¬ sets. A memory representation of this binary-coded If necessary, structure document BCS is loaded into the object ^interface Ol.

The processing processes XP1, XP2, XP3 perform a processing operation on binary-coded structure documents, for example in the format EXI (Efficient XML Interchange). This processing includes in particular a durchsu ^¬ chen, filter, manipulate, transform, and query of data of the data contained in binary coded structure documents. Examples include machining processes - how in the trade Bored often, "also referred to as" EXI processors - possibility based on the query or transfor ^¬ mationssprachen XPath, XQuery, XSLT, etc. Due to the precisely explained below according to the invention of a processing of binary-coded structure document BCS in conjunction with the binary-coded index document BCI, the invention advantageously allows use of the transformation language XQuery instead of the limited query language XPath. The possibilities of the transformation language XQuery go beyond a mere query. It is a turing-complete language with extensive transformation capabilities that go far beyond XPath's query and filtering capabilities. In contrast, the query language XPath provides only part ^¬ extensive or subset of the capabilities of XQuery.

The following description of the figures follows with further reference to the functional units of FIG. 1. Identical reference symbols in different figures here represent identical functional units.

Fig. 2 shows a schematic representation of a structure of the index document BCI in comparison with an underlying in a text-based description language, here XML, constructed structural document SD.

The exemplary structure document SD shown in structure blocks in the left half of the picture is based on the following XML code:

<? xml version = "l .0" encoding = "UTF-8"?>

<body> Do not forget it! </ body>

</ Note>

<subj ect> shopping list </ subj ect>

<body> milk, honey </ body>

</ Note>

</ notebook> This is an XML representation of a Notizblo ^¬ ckes containing two notes. The notes as the no ^¬ tizblock themselves are referenced to a respective date in an ISO 6601 format of the form YYYY-MM-DD (year-month-day) as an attribute. The first note contains another attribute, an "EXI" category. In addition, each has no ^¬ tiz about a subject or "Subject" and the actual content or "Body".

In the left-hand half of FIG. 2, the structural document SD containing the XML code shown above is shown in individual structure blocks. In the right-hand half of the figure, a schematic representation of a binary-coded index document BCI formed from the structure document SD is shown, to which, according to the invention, a structural subset of the structure is shown. document SD. In the drawing is recognizable ^¬ bar, that the structure blocks of the structure document SD jewei ^¬ celled structure blocks of binary coded index document BCI correspond as angedeu- by broken reference arrows tet.

According to the usual EXI coding measures, compact identification numbers or "identifiers" are provided in the binary-coded index document BCI, which replace a character string or "string" contained in the structure document SD. If a character string occurs repeatedly in the structure document SD, then in the EXI coding this character string is referenced as a qualified name or "qname" by the corresponding identifier. For this a maintenance of - not shown - resource tables or »EXI String Tables« is provided. In the present example, the identifier 5 in the binary-coded index document BCI stands for the qualified name "notebook", the identifier 4 for "note", the identifier 6 for "subject" and the identifier 1 for the qualified name "body".

The binary-coded index document BCI is based on a subset of the XML structure document shown which, according to the embodiment of the invention shown here, is restricted to the structural index but does not contain the actual values, eg the content "shopping list" of the subject subject. The binary-coded index document BCI ent ^¬ therefore keeps te a binary representation of scarce hierarchical data structure of the elementary structure defined in the blocks, without the zughörigen content.

The structural part of the circumference of the present invention Strukturdo ^¬ kuments BCI comprises in alternative embodiments, in addition to the scarce hierarchical data structure if necessary also contents, which are often needed eg with inquiries.

The graphical representation in structure blocks could easily be done in a tree structure without affecting the structure of the binary-coded index document BCI itself.

An inquiry - which constitutes a special case of a machining operation - can be used in majority of cases already with the binary-coded index document BCI itself answered the ^¬. Such requests include, in particular, queries on a hierarchical context, on a number or on specific categories of elements to be determined. Further processing operations, ie modifications or transformations of structural data, can in most cases only be carried out in cooperation with the binary-coded index document BCI. Is such a machining ^¬ processing operation associated with a modification in the structure of the binary coded index BCI document, the Modifika ^¬ tion is performed by the object interface Ol also in binary coded document structure BCS. In this way, a consistency of the binary-coded index document BCI-as a structural image of the binary-coded structure document BCS-with the binary-coded structure document BCS is ensured. To carry out the modification by the object interface Ol in binary coded structure document BCS Ask ^¬ elements of the binary-coded document structure by the ^¬ Whether jektschnittstelle oil are loaded into memory. The access to values in the binary-coded structural document takes place via the identification number already known from the binary-coded index document. This measure allows a ra ^¬ rule and random access to the values to be modified. The object interface Ol is designed especially according to the "Docu- ment Object Model" or DOM which a Datenzu ^¬ grabbed for example in the form of a programming interface or API (Application Programming Interface) allows.

The object interface oil is from the perspective of Bearbeitungspro ^¬ processes XP1, XP2, XP3 so approachable than they would completeness, ^¬ ended binary coded loaded structure document BCS. However, the object interface Ol holds initially only the structure of the information provided in binary coded index document BCI and invite by using structural information provided in binary coded index document BCI if necessary parts of binärko ^¬-founded structure document BCS.

In the following two exemplary Bearbeitungsoperatio ^¬ nen are explained with a description of the involvement of the binary-coded index document BCI.

The first question was the question of how many notes the notebook has. Such a request is made, for example, by one of the processing processes XP1, XP2, XP3 on the basis of the query languages XPath or XQuery, for example in the query syntax: "count (/ notebook / note /)". This to ask ^¬ can be answered with the exclusive participation of the binary-coded index document BCI by all local elements with the identification number 4 - representing the string "note" - are counted.

Subject of a second inquiry is the question of how many no ^¬ corporate rank with the category "EXI" are applied. A Derar ^¬ term request is made, for example, in the following query syntax "count (/ notebook / note [@ category = ΈΧΙ '])" · This request would Follow and resolve the two ID numbers of ID 4 to verify that the category attribute exists and has the value EXI. The dissolution takes place with participation of the - not shown - resource table or »EXI String Table«. In order to check whether one of the »category« attributes has the value »EXI«, however, unlike the above query, the binary-coded structure document BCS is to be used, since the contents of the attributes are not contained in the structural subset of the binary-coded index document BCI. The object fragment Ol therefore loads the corresponding fragment of the binary-coded structure document BCS and performs the check. The binary-coded document structure BCS as an actual fully ^¬ constant information carrier document is represented by means of the binä ^¬ ren format EXI one hand compact and has index jump labels, which permit a random access to each element. To realize a random access, for example, the »seifContained« property of elements provided in the EXI specifications is used.

Further, a use of the known substitution of element names by a corresponding identification number also results in binary coded index document BCI cause the bi ^¬ närkodierte Index document BCI compared to its zugrun ^¬ deliegenden structure document SD is kept very slim. The restriction according to the invention to a structural subset with largely omission of values leads to an even greater streamlining of this Xlndex, which can be stored with the inventive measures completely in a working memory. Must to achieve this random access to seifContained elements of the binary-coded structure document BCS single ^¬ Lich the relevant element - which is to be understood as requiring as needed fragment in this context - the binary-coded structure document BCS are loaded for processing by the object interface Ol, ie in front of not the complete structure document BCS or a more extensive part of the structure of the document BCS ^¬ finally any forward the relevant element ^¬ some manner and downstream outsourced document content.

According to one embodiment of the invention, it is provided that the binary-coded index document contains not only the structural index but also further, in particular frequently required contents and values. A processing of these frequently used content and values thus does not require repeated charge ent ^¬ speaking fragments of the binary-coded structure document BCS by the object interface oil. Instead, application can be processed this frequently used content and values with the exclusive participation of the binary-coded index document BCI.

According to one embodiment of the invention, provision is made for an inline entry "<inline />" to be provided in the structure-coded document SD underlying the binary-coded index document BCI, which indicates that a so-marked element is not a reference but a copy of the value from the binary-coded structure document BCS contains. One marked with an inline entry "<inline />"

Element in the generated therefrom binary coded index document BCI underlying structure document SD means that the element so designated - more precisely, a Binärrepräsen ^¬ tation of marked element - in the binary-coded Index document BCI is taken and is therefore immediately accessible for machining operations without the fragment by the object interface oil from the binary coding ^¬ th structure document BCS advertising loaded into memory to need.

Inline entries prove advantageous to accelerate the Anfra ^¬ give processing at the expense of the size of the binary-coded index document BCI. Many requests also require specific values for evaluation, eg for evaluating filter predicates. If these values are present in inline elements of the binary-coded index document BCI, access to the binary-coded structure document BCS does not have to be made beyond the index. This can be useful for values that are frequently referenced in query selection predicates.

According to another embodiment of the invention, a ghost entry "<ghost />" in the binary coded index document BCI underlying structure document SD watch before ^¬ merely indicating the existence of an element but beyond that no further information, particularly any content and / or values. If required, the content itself can be referenced to the binary-coded structure document BCS or to another

- Not shown - binary-coded index document be ^¬ read.

A labeled with a ghost entry "<ghost />" element in the generated therefrom binary coded index document BCI underlying structure document SD causes the element so designated is not included in the binary coded Indexdo ^¬ kument BCI, but only a Ghost Entry for this element. The binary-coded index document ment BCI registered for this element Ghost record label ^¬ draws only the existence of this element, but not the element or its contents. Should this element required by a processing operation, it must, if required by the object interface oil from the binary ^¬ coded Structure document BCS be loaded into memory.

Ghosted item enables a slender binary coded index document BCI even with a large underlying structure document SD. This measure allows partitioning of one or more binary-coded index documents and paging of those portions of the index from main memory that are likely to be rarely needed.

The following again refers to the above XML representation of the scratch pad. A EXI representation of the bi ^¬ närkodierten Index BCI document has the following structure:

<? xml version = "l .0" encoding = "UTF-8"?>

<x: element xmlns: xsi = "http: // w .w3.org / 2001 / XMLSchema-instance" xsi: schemalocation = "http://www.siemens.com/XIndex xindex.xsd xmlns: x =" http : // www. Siemens. com / xIndex ">

<x: offsetFirstChild> 4 </ x: offsetFirstChild>

<x: element>

<x: offset irstChild> 8 </ x: offset irstChild>

<x: element>

<x: offsetNextSiblinq> 6 </ x: offsetNextSibl

</ x: element>

<x: element> <Body />

<x: offsetNextParentSibling> 2 K / x: offsetNextParentSibling> </ x: element>

</ x: element>

<x: element>

<x: offset irstChild> 4 </ x: offset irstChild>

<x: element>

<x: offsetNextSiblinq> 16 </ x: offsetNextSiblinq>

</ x: element>

<x: element>

<x: offsetNextParentSibling> 15 </ x: offsetNextParentSibling> </ x: element>

</ x: element>

The binary-coded index document BCI has a jump label

in the actual binary coded structure document BCS before ^¬ preferably via a 'local' offset to the next

Element. Due to the different type of the next one

Element is shown in the EXI representation of the

binary-coded index document BCI for better understanding

namely between <offsetNextSibling />,

<offsetNextParentSibling /> and <offsetFirstChild />.

There are exactly one of each of these elements for each element

Offsets and technically the name distinction of the different types is not necessary. Alternatively, the

Offset therefore also generically designated, e.g. when

<offsetNextElernent InDocumentOrder />. In contrast to the local offset described above, an absolute offset would have the advantage that the number of bytes to be skipped in the binary-coded structure document BCS in order to be able to read in the complete relevant element would have to be determined in one step. In the case of struc ^¬ tural or substantive modifications to the data contained in the binary-coded document structure BCS but all subsequent Elementin- would have to be updated dices possibly at an absolute offset. When using a local offset, such modifications of an item merely result in an update in the overlying hierarchy.

Local offsets are advantageously used to respond to a request from one of the editing processes

XP1, XP2, XP3 skip a non-relevant part to Kings ^¬ nen. For example, a request for

/ Notebook / note / body

cause the <subject> element before body to be irrelevant and to be skipped by the offset.

Another example would be a request according to

/ Notebook / editor

which is looking for an element "editor" that does not actually exist. As soon as the request encounters a <notebook> element, it is skipped directly because it contains no <editor> element and is therefore irrelevant. The length of an element to be skipped is calculated according to the following rule:

1. Initialize length with 0

2. Add length with <offsetFirstChild> if available and jump to firstChild index OR if not available

Add either <offsetNextSibling> OR

<offsetNextParentSibling> (one is present) and jump to the said index

3. Continue with step 2. until

a. if element has children (started with <offsetFirstChild>) until every child with

<offsetNextParentSibling> has completed;

b. if element has no children then includes

<offsetNextSibling> the length.

The following table summarizes the offsets.

In it means in the column heads:

AO: absolute offset (absoulute offset)

L: Length in bytes

OFC: Offset to first child (offsetFirstChild) ONS: Offset to next sibling (offsetNextSibling) ONP: Offset to next parent (offsetNextParentSibling)

In the table, a start tag of an EXI element with start element or SE, the corresponding event or event on the end tag is referred to as end element or EE. The attributes of a start tag are entered with AT. The event followed by the data of an element is called Character or CH.

A calculation of the length for notebook [l] gives 74, ie 4 + 8 + 6 + 21 + 4 + 16 + 15

A calculation of the length for note [l] gives 35, ie

8 + 6 + 21

A calculation of the length for note [2] gives 35: 4 + 16 + 15 A calculation of the length for note [2] / subj ect [1] yields

16:16

The execution of machining operations according to the invention on the binary-coded index document BCI, which is also referred to as XIndex, has the following advantages:

- The method allows faster query processing by certain queries can already be processed by means of Indexdocu ^¬ management BCI;

Significantly reduced memory usage during processing of requests, both on resource-limited devices and on other computer systems, all the way to servers;

- The method permits in addition to a request ification of the data contained in the document index BCI which stam from the underlying basis constructed in a text-based description language document structure SD ^¬ men also an MO; - The process provides as a result already present ^¬ the data in more efficient form, that is by means of the format EXI;

- The process allows the use of Standardtechnolo ^¬ technologies such as the transformation language XQuery on resource-limited devices such as embedded systems, which in contrast to the XPath query language allows a transformation of data on a pure queries or filters.

Claims

claims

1. A method for processing a binary-coded structure document in a resource-limited device in particular, wherein the binary-coded structure document (BCS) based in a text-based description language Strukturdoku ^¬ ment (SD) is based and wherein a binary coded index document (BCI) a structural subset of the structure document (SD ) underlying;

comprising the steps:

- depositing an in-memory representation of the binary-coded index document (BCI) in a working memory of the Ge ^¬ Raets;

Depositing the binary-coded structure document (BCS) in a background memory associated with the device;

Performing a processing operation by at least one processing process (XP1, XP2, XP3) with access to the memory representation of the binary-coded index document (BCI) via an object interface (Ol), possibly fragments of the binary-coded structure document

(BCS) are loaded by the object interface (Ol) in the working ^¬ memory.

2. The method according to claim 1, characterized in that the binary-coded index document (BCI) and / or the bi ^¬ närkodierte structure document (BCS) are present in an efficient XML interchange format or EXI format.

3. The method according to any one of the preceding claims, character- ized in that the text-based description ^¬ Language is an Extensible Markup Language or XML.

4. The method according to any one of the preceding claims, characterized in that in the binary-coded index document (BCI) elements due to an identification of a corre ^¬ sponding element in the underlying structure document (SD) are immediately accessible for editing operations accessible in the binary coded index document (BCI).

5. The method according to any one of the preceding claims 1 to 3, characterized in that in the binary-coded index ^¬ document (BCI) elements due to an identification of a corresponding element in the underlying structure document (SD) are excluded in the binary coded index document (BCI) and that the elements in Requirement from the object ^¬ interface (Ol) from the binary-coded structure document (BCS) removed and loaded into the main memory.

6. Computer program product with means for performing the method according to one of the preceding claims, when the computer program product is brought to a particular res ^¬ resource-limited device for execution.

7. Device for in particular resource-limited processing of a binary-coded structure document, comprising

a background memory assigned to the device for storing the binary-coded structure document, the structure document being based on a structure document constructed in a text-based writing language;

with a working memory for storing a binary-coded index document on the basis of a structural partial extent of the structural document;

- process with an object interface for an access to the binary-coded index document, in performing Be ^¬ processing operation by at least one processing, the object interface established for a BE ^¬ can as loading fragments of the binary-coded structure of the document in the main memory.